Anomalous Vapor Sensor Response of a Fluorinated Alkylthiol

Feb 25, 2009 - Materials and Transducers Toward Selective Wireless Gas Sensing. Radislav A. Potyrailo , Cheryl Surman , Nandini Nagraj , and Andrew ...
1 downloads 0 Views 2MB Size
pubs.acs.org/Langmuir © 2009 American Chemical Society

Anomalous Vapor Sensor Response of a Fluorinated Alkylthiol-Protected Gold Nanoparticle Film Jisun Im, Amol Chandekar, and James E. Whitten* Department of Chemistry and Center for High-Rate Nanomanufacturing, The University of Massachusetts Lowell, Lowell, Massachusetts 01854 Received January 3, 2009. Revised Manuscript Received February 5, 2009 Monolayer-protected gold nanoparticle films generally swell and increase their electrical resistance when exposed to organic vapors. Films of gold nanoparticles protected by 1H,1H,2H,2H-perfluorodecanethiol (PFDT) exhibit an anomalous response in which the resistance decreases for all vapors investigated. Electron microscopy illustrates that the PFDT-functionalized gold nanoparticles are hexagonally ordered with an interparticle separation of 3 nm. Quartz crystal microbalance measurements confirm substantial mass uptake, but the relatively large interparticle separation and insulating properties of the gold particles lead to a porous film whose electrical resistance is strongly influenced by changes in the relative permittivity and reversible, vaporinduced changes in film morphology.

Introduction A variety of chemiresistive sensors that include phthalocyanine,1-3 conjugated polymer,4,5 and carbon nanotube6,7 thin films have been developed to detect a range of analytes. In 1998, Wohltjen and Snow8 demonstrated a chemical vapor sensor based on thiol-protected gold nanoparticles. The mechanism of charge transport involves a combination of electron tunneling and hopping. The sorption of analyte organic vapor molecules generally causes swelling of the film, an increase in interparticle distance, and a concomitant increase in electrical resistance. The concentration of absorbed organic molecules is dictated by the partition coefficient of the vapor between the gas and film,9 which depends on the activity coefficient for dissolution of the vapor into the film and the vapor pressure of the analyte. In principle, it is possible to develop a selective sensor by measuring the resistance changes of two or more different films. The electrical conductivity of gold nanoparticle films is determined by several factors, including the interparticle separation and the relative permittivity,10,11 which can change as a result of the sorption of gas molecules. For polar analytes, *Corresponding author. Tel: (978) 934-3666. Fax: (978) 934-3013. E-mail: [email protected]. (1) Bohrer, F. I.; Colesniuc, C. N.; Park, J.; Schuller, I. K.; Kummel, A. C.; Trogler, W. C. J. Am. Chem. Soc. 2008,::130, 3712–3713. :: :: :: (2) Zhou, R.; Josse, F.; Gopel, W.; Ozturk, Z. Z.; Bekaroglu, O. Appl. Organometal. Chem. 1996, 10, 557–577. (3) Cook, M. J. J. Mater. Chem. 1996, 6, 677–689. (4) McQuade, D. T.; Pullen, A. E.; Swager, T. M. Chem. Rev. 2000, 100, 2537–2574. (5) Toal, S. J.; Trogler, W. C. J. Mater. Chem. 2006, 16, 2871–2883. (6) Kauffman, D. R.; Star, A. Angew. Chem., Int. Ed. 2008, 47, 6550–6570. (7) Zhang, T.; Mubeen, S.; Myung, N. V.; Deshusses, M. A. Nanotechnology 2008, 19, 332001:1–14. (8) Wohltjen, H.; Snow, A. W. Anal. Chem. 1998, 70, 2856–2859. (9) (a) Snow, A. W.; Wohltjen, H.; Jarvis, N. L. In Defense Applications of Nanomaterials; Miziolek, A. W., Karma, S. P., Mauro, J. M., Vaia, R. A., Eds.; ACS Symposium Series 891; American Chemical Society: Washington, DC, 2005; pp 31-45. (b) An excellent review by Snow, A. W.; Wohltjen, H.; Jarvis, N. L. is available at http://www.nrl.navy.mil/content.php?P=02REVIEW45. (10) Neugebauer, C. A.; Webb, M. B. J. Appl. Phys. 1962, 33, 74–82. (11) Zabet-Khosousi, A.; Dhirani, A.-A. Chem. Rev. 2008, 108, 4072– 4124.

4288

DOI: 10.1021/la900016u

a decrease in resistance may occur,8,9,12,13 especially at high concentrations, as a result of the modification of the permittivity of the medium between the gold particles. In this letter, we report that films of gold nanoparticles protected by a fluorinated alkanethiol monolayer exhibit unusual behavior compared with those protected by other thiol ligands. Because of the large separation between adjacent gold nanoparticles, apparently caused by their bulky, rigid helical ligands that are not easily interleaved by those of neighboring particles, electron tunneling and hopping between adjacent gold cores is not the major conduction mechanism. The conductivity of the film is instead strongly influenced by the relative permittivity of the dielectric medium between the gold cores, which is modified by analyte absorption. For all vapors tested, the resistance decreases when the gold particle film is exposed to analytes. The possible contribution of reversible, vaporinduced changes in film morphology is also discussed.

Experimental Section Materials and Syntheses. 1H,1H,2H,2H-Perfluorodecanethiol (PFDT) was obtained from Oakwood Products. Other solvents and reagents, including 16-mercaptohexadecanoic acid (MHDA), were used as received from Aldrich. The synthesis of the alkanethiol-protected gold nanoparticles was performed by modifying Brust’s method.14 Gold chloride (HAuCl4) in the aqueous phase was transferred by a phase-transfer reagent, (C8H17)4NBr, into a toluene alkanethiol solution using a 3:1 AuCl4 /thiol molar ratio. Aqueous sodium borohydride (NaBH4) was then added dropwise while stirring, and the mixture was stirred vigorously for 12 h under ambient conditions. The dark, violet-colored organic phase was separated from the aqueous phase, rotary evaporated to 3 mL, diluted to 300 mL with ethanol, and stored overnight at -70 °C. The product was washed three times with ethanol, separated by centrifugation, and dried under vacuum. Once dried, the (12) (a) Evans, S. D.; Johnson, S. R.; Cheng, Y. L.; Shen, T. J. Mater. Chem. 2000, 10, 183–188. (b) Zhang, H.-L.; Evans, S. D.; Henderson, J. R.; Miles, R. E.; Shen, T.-H. Nanotechnology 2002, 13, 439–444. (13) Ahn, A.; Chandekar, A.; Kang, B.; Sung, C.; Whitten, J. E. Chem. Mater. 2004, 16, 3274–3278. (14) Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman, R. J. Chem. Soc., Chem. Commun. 1994, 7, 801–802.

Published on Web 2/25/2009

Langmuir 2009, 25(8), 4288–4292

Letter PFDT-protected gold nanoparticles (PFDT-AuNPs) were found to be soluble only in nonpolar fluorinated solvents such as hexafluorobenzene. Dried MHDA-protected gold nanoparticles (MHDA-AuNPs) were soluble in methanol. Characterization. X-ray photoelectron spectroscopy measurements of gold nanoparticle films spun cast from 1 mg/mL solutions onto indium tin oxide substrates were performed in a VG ESCALAB MK II photoelectron spectrometer equipped with a Mg KR X-ray source. Absorbance spectra were acquired using a Lambda 1806 UV-vis spectrometer. Transmission electron microscopy (TEM) images of 2 mg/mL solutions drop cast onto carbon-coated copper grids were obtained using a Philips EM 400t microscope and an accelerating voltage of 100 kV. Electrical Conductivity and Mass Uptake. Gold nanoparticle films were prepared by drop casting 2 mg/mL solutions onto interdigitated array (IDA) microelectrodes (M1450110, Microsensor Systems, Inc.). These consist of 50 pairs of gold electrodes with the following dimensions: 15 μm electrode width, 15 μm spacing, 4800 μm overlapping length, and 1500 A˚ electrode thickness. Conductivity, σ, is calculated from d I σ ¼ ð1Þ ð2n -1Þ  L  h  V where d is the electrode spacing, I is the current, n is the number of electrodes, L is their overlapping length, h is the film thickness, and V is the voltage difference between the electrode pairs. This equation is valid when the thickness of the film does not exceed that of the gold electrodes, as in the present study. For the results reported here, 2.0 dc voltage was used, and the thickness of the gold nanoparticle film was measured by atomic force microscopy. Ohmic behavior was observed in this voltage range. To measure the sensing properties of a gold nanoparticlecovered IDA electrode, it was exposed to vapor while measuring its electrical resistance. Mass uptake of the gold nanoparticle films upon vapor exposure was performed with gold-coated, 5 MHz AT-cut quartz crystal microbalance (QCM) electrodes onto which gold nanoparticle films had been deposited using the same concentration (2 mg/mL) as for the vapor sensors. The measured frequency change was converted to mass change using the Sauerbrey equation,15 as detailed in the Supporting Information.

Results and Discussion XPS of the monolayer-protected gold nanoparticle films confirmed their successful syntheses. In the case of PFDTAuNPs, an intense F 1s peak at 689.0 eV was observed, along with three C 1s components at 285.2, 291.6, and 293.8 eV corresponding to -CH2-, -CF2-, and -CF3, respectively. An S 2p peak at 162.8 eV confirmed thiol-Au bonding.16 The PFDT-AuNP in hexafluorobenzene and MHDA-AuNP in methanol solutions exhibited UV-vis spectroscopy absorbance peaks at 504 and 512 nm, respectively. These correspond to the gold plasmon band, confirming that these solutions contain isolated nanometer-sized gold particles. Figure 1a displays the sensor response of an MHDA-AuNP film that exhibits behavior typical of monolayer-protected gold nanoparticle films. Figure 1b shows the response of a PFDTAuNP film to the same analytes. In the case of the MHDAAuNP film, the sensitivity order, at moderate to high concentrations, is toluene > water > hexafluorobenzene > hexane = methanol > chloroform, with the electrical resistance increasing upon vapor exposure, consistent with a swelling mechanism. For the PFDT-AuNP film, the corresponding order is hexafluorobenzene > methanol > hexane > chloro-

form > toluene = water, but in all cases, the electrical resistance decreases and the conductivity increases. This behavior indicates that swelling is not the dominant mechanism. An expansion of the low-concentration region of Figure 1b (Supporting Information) indicates essentially no response of the PFDT-AuNP film to chloroform, water, or hexane at concentrations below 50 000 ppm and a sensitivity order of toluene > methanol > hexafluorobenzene. An alternative approach8 to displaying the data in Figure 1, which attempts to decouple the effect of analyte vapor pressure from analyte-particle interaction, is also provided in the Supporting Information section. To understand the operative mechanism for the resistance decrease of the PFDT-AuNP film, its mass uptake upon vapor exposure was investigated. As shown in Figure 2a, hexafluorobenzene exhibits the greatest absorption into the film (4.20 μg/cm2 at 100 000 ppm), and methanol is the least absorbed, even at high concentration (0.55 μg/cm2 at 150 000 ppm). These results demonstrate that significant absorption occurs, even though the relative resistance does not increase. Figure 2b contains the results from combining data from Figures 1b and 2a and displays relative resistance changes versus mass uptake. An interesting aspect of this graph is that the vapors exhibit significant mass uptake thresholds below which only minimal decreases in resistance are observed. Furthermore, the resistance changes exhibit at least two regimes, becoming somewhat linear for large mass uptakes, as especially evident for methanol and hexane. This combined graph illustrates that whereas the PFDT-AuNP sensor is most sensitive to methanol and hexafluorobenzene, two molecules having very different polarities, when resistance is plotted versus concentration the film is significantly more responsive to methanol, per absorbed microgram/cm2 of analyte, compared to hexafluorobenzene. Figure 3 shows TEM images of the MHDA and PFDT gold nanoparticle films. The mean core size of the MHDA-AuNPs is ca. 5 nm, with an average edge-to-edge (interparticle) distance of 1 to 2 nm. In the case of the PFDT-AuNPs, the mean core size is ca. 2.5 nm, with an average interparticle distance of ca. 3 nm. The different core size is likely not the origin of the anomalous behavior because both of these core diameters are in the range typically used for gas sensors.9 An interesting aspect of the PFDT-TEM image (Figure 3b) is that the monolayer-protected gold particles within the film are hexagonally ordered. Film thicknesses of 60 and 80 nm and conductivities of 1.44  10-5 and 3.35  10-9 S/cm have been measured for the MHDA-AuNP and PFDT-AuNP films, respectively. The conductivity depends on the particle size, the interparticle distance, the dielectric properties of the organic ligands, and the interstitial regions between the particles. Zabet-Khosousi and Dhirani11 have reviewed the properties of monolayerprotected gold nanoparticle films. Typical conductivities for alkylthiol ligands are in the range of 10-2 to 10-6 S/cm, with the value generally decreasing as the length of the thiol chain increases for a particular core size. In the present study, the conductivity of the MHDA-protected AuNP film is in the commonly observed range, whereas that of the PFDTprotected AuNP film is orders of magnitude lower. The conductivity (σ) of the nanoparticle gold film at temperature T is described10,11 by σ ¼ expð -βsÞ expð -Ec =kB TÞ

(15) Sauerbrey, G. Z. Phys. 1959, 155, 206–222. (16) Nuzzo, R. G.; Fusco, F. A.; Allara, D. L. J. Am. Chem. Soc. 1987, 109, 2358–2368.

Langmuir 2009, 25(8), 4288–4292

ð2Þ

where β is a decay constant related to the probability of an electron tunneling from one gold particle to another (typically DOI: 10.1021/la900016u

4289

Letter

Figure 1. Sensor responses of (a) MHDA-AuNP and (b) PFDTAuNP films to toluene, hexafluorobenzene, hexane, methanol, chloroform, and water vapors, using nitrogen as the carrier gas. The sensor response is presented in terms of the change in resistance divided by the initial resistance. 1 A˚-1), s is the interparticle distance, Ec is the activation energy, and kB is the Boltzmann constant. The activation energy is given by Ec ¼ e2 =4πεr εo r

ð3Þ

where e is the fundamental charge, εo is the vacuum permittivity constant, r is the gold nanoparticle radius, and εr is the dielectric constant of the film. There are two major reasons for the lower conductivity of the PFDT-AuNP films. First, the mean interparticle separation is 3 nm. Typical separation values11 for protected AuNPs are 1 to 2 nm, as observed for the MHDA-AuNP film. The larger separation in the case of the PFDT-AuNPs leads to decreased probability of tunneling between particles, which decays exponentially with distance. The second major factor is the larger tunneling barrier presented by the PFDT ligands on the gold particles. The breakdown voltage of a PFDT monolayer self-assembled on a gold 4290

DOI: 10.1021/la900016u

Figure 2. (a) Mass uptake of a PFDT-AuNP-coated 5 MHz quartz crystal microbalance upon exposure to toluene, hexafluorobenzene, hexane, methanol, chloroform, and water vapors. (b) Relative resistance change versus mass uptake. surface has been measured by Whitesides and colleagues17 to be 2.2 ( 0.4 V. For reference, a corresponding value of 1.3 ( 0.3 V has been measured for 1-dodecanethiol.17 The combination of the larger interparticle separation and larger tunneling barrier leads to the significantly lower conductivity of the PFDT-AuNP film. Other researchers have previously synthesized PFDTAuNPs. However, they did not study their sensor properties. Murray and co-workers18 recently demonstrated a new synthetic ligand replacement strategy and compared the resulting gold core sizes and polydispersities. Yonezawa and colleagues,19 using the Brust method, measured an interparticle spacing of 2.4 nm and observed hexagonal packing of the gold particles in drop-cast films. The TEM image shown in (17) Haag, R.; Rampi, M. A.; Holmlin, R. E.; Whitesides, G. M. J. Am. Chem. Soc. 1999, 121, 7895–7906. (18) Dass, A.; Guo, R.; Tracy, J. B.; Balasubramanian, R.; Douglas, A. D.; Murray, R. W. Langmuir 2008, 24, 310–315. (19) Yonezawa, T.; Onoue, S.; Kimizuka, N. Langmuir 2001, 17, 2291–2293.

Langmuir 2009, 25(8), 4288–4292

Letter

Figure 3. TEM micrographs of drop-cast films of (a) MHDA- and (b) PFDT-protected gold nanoparticles. The insets show an expansion of the micrographs, with the dimension scale bar equal to 20 nm. Figure 3a is consistent with their observations, although our measured interparticle distance is greater. The authors of ref 19 attributed the formation of the uniform, hexagonally packed nanoparticle film to the low surface tension of the fluorinated monolayer, which causes the particles to spread over a large area. For MHDA-AuNPs and other protected AuNPs that we have studied, the small interparticle separation suggests that the thiol chains overlap considerably with each other. In the case of the TEM image of the PFDT-AuNPs presented in Figure 3, the interparticle spacing is ca. 30 A˚. The molecular length of PFDT is 14.6 A˚.20 The fact that the interparticle (20) Tsao, M.-W.; Hoffmann, C. L.; Rabolt, J. F.; Johnson, H. E.; Castner, D. G.; Erdelen, C.; Ringsdorf, H. Langmuir 1997, 13, 4317–4322.

Langmuir 2009, 25(8), 4288–4292

distance is approximately twice that of the PFDT length suggests that the gold particles are packed such that the tails of the PFDT molecules just touch each other, although this assumes an extended conformation. Perfluorocarbon chains are believed to assume a helical conformation20 with a van der Waals diameter of 5.6 A˚21 and to be stiff compared to hydrocarbon chains.22 The corresponding diameter of a hydrocarbon chain is 4.2 A˚.23 The larger radius compared to that of a hydrocarbon chain and a higher barrier to rotation about the C-C axes24 lead to a more rigid monolayer that is difficult to be interleaved by ligands from neighboring particles. Because the formation of a helical configuration is expected to reduce the length of the PFDT molecule, it is likely that the large interparticle separation results from a combination of low surface tension and ligand rigidity. The anomalous sensor response of the PFDT-protected gold nanoparticle film originates from the relatively large spacing between neighboring PFDT-AuNPs. Because the gold particles are spread out, this leads to a film that is porous and that can accommodate high concentrations of analytes. The conductivity of the film increases, at least partially, because of the increased dielectric constant of the medium between the gold cores upon vapor sorption, and the resistance of the film decreases. Although the present study is the first to note the unusual sensor behavior of a fluorinated alkanethiol-protected gold nanoparticle film, Joseph and coworkers25,26 reported increased conductivity of networked gold nanoparticle chemiresistors functionalized with [4]-staffane-3,3000 -dithiol upon exposure to a variety of solvents, including toluene. They attributed the behavior to the suppression of swelling due to the rigid linkers, with vapor sorption leading to increases in the permittivity of the film. In the present study, we propose that a similar phenomenon is observed except that whereas swelling may occur its effect is minimized because the gold particles are far apart and insulated from each other prior to vapor sorption. Attempts to correlate the experimentally measured decrease in resistance quantitatively with the dielectric constant of the vapor and the number of absorbed molecules do not give satisfactory results (Supporting Information). This suggests that several factors are responsible for the observed behavior of the PFDT-AuNP film. The decrease in resistance, caused by the relative permittivity increase, may to some extent be offset by film swelling. Additionally, vapor-induced changes in film morphology may occur. For example, it has recently been hypothesized that the observed decrease in the resistance of thin films of layers of gold nanoparticles linked by dodecanedithiol may be due to a decrease in inter-island separation caused by swelling along the film normal.27 In the present study, it is possible that the PFDT-AuNPs form electrically isolated islands and that vapor-induced changes (21) Lenk, T. J.; Hallmark, V. M.; Hoffmann, C. L.; Rabolt, J. F.; Castner, D. G.; Erdelen, C.; Ringsdorf, H. Langmuir 1994, 10, 4610–4617. (22) Snow, A. W.; Buckley, L. J. Macromolecules 1997, 30, 394–405. (23) Gentilini, C.; Evangelista, F.; Rudolf, P.; Franchi, P.; Lucarini, M.; Pasquato, L. J. Am. Chem. Soc. 2008, 130, 15678–15682. (24) Bates, T. W. In Fluoropolymers; Wall, L. A., Ed.; John Wiley & Sons: New York, 1972; pp 451-474. (25) Joseph, Y.; Besnard, I.; Rosenberger, M.; Guse, B.; Nothofer, H. G.; Wessels, J. M.; Wild, U.; Gericke, A. K.; Su, D.; Schlogl, R.; Yasuda, A.; Vossmeyer, T. J. Phys. Chem. B 2003, 107, 7406–7413. (26) Joseph, Y.; Peic, A.; Chen, X.; Michl, J.; Vossmeyer, T.; Yasuda, A. J. Phys. Chem. C 2007, 111, 12855–12859. (27) Joseph, Y.; Guse, B.; Vossmeyer, T.; Yasuda, A. J. Phys. Chem. C 2008, 112, 12507–12514.

DOI: 10.1021/la900016u

4291

Letter

in film morphology push the islands closer together or cause islands of nanoparticles to spread out, forming percolation pathways. Such behavior would be consistent with the observed mass uptake thresholds. Further experiments are underway to understand the details of the PFDT-AuNP sensor behavior quantitatively.

Conclusions PFDT-AuNP sensor films exhibit an unexpected vapor response in which their conductivity increases upon exposure to analytes. The rigid, helical nature of the perfluorocarbon chains leads to minimal interleaving of ligands from nearby particles, resulting in large interparticle separation. As a voltage is applied to the sensor, charging of the gold nanoparticles occurs. Analyte absorption into the film causes the relative permittivity of the medium surrounding the nanoparticles to increase (as dictated by eqs 2 and 3), and electrons tunnel between the nanoparticles, leading to an observed decrease in film resistance. Reversible changes in film morphology, including inter-island separation distance, may also play a role.

4292

DOI: 10.1021/la900016u

It is important to consider the utility of vapor sensors based on PFDT-AuNP films, or other films, that respond to analytes mainly by permittivity changes instead of swelling. Whereas these may not exhibit the selectivity of thiolprotected gold particles that change their electrical resistance as a result of swelling, they may find niche applications as components of sensor systems that respond to more than one property of an analyte mixture and may be advantageous in terms of selectivity. Acknowledgment. This work was supported by a multisensor grant from the Army Research Laboratory to the University of Massachusetts Lowell. We acknowledge helpful suggestions from Dr. Arthur Snow at the Naval Research Laboratory. Supporting Information Available: XPS spectra of the fluorinated gold nanoparticle films, sensor curves in which the responses are normalized to the vapor pressures of the analytes, and additional comments. This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2009, 25(8), 4288–4292